What types of referral fraud does this skill help detect?

It monitors for common abuse signals including IP/device matching, velocity spikes, duplicate accounts, and geographical mismatches.

How does it help with compliance reporting?

It includes templates for fraud monitoring dashboards and investigation logs, making it easier to report status and risks to stakeholders like legal and finance.

Is this skill only for technical developers?

While designed for Claude Code, it provides templates and runbooks useful for marketing operations, finance, and legal teams to manage program integrity.

Does it support manual review processes?

Yes, the framework explicitly combines automated checks with random manual audits to ensure accuracy and reduce false positives.

Can this skill automate the blocking of fraudulent users?

Yes, it includes workflow automation patterns to auto-flag suspicious activity or pause reward payouts until a manual review is completed.

Referral Fraud Detection

Name: Referral Fraud Detection
Author: gtmagents

bygtmagents

•

マーケティングオートメーション

Protects growth campaigns by monitoring, investigating, and preventing referral program abuse and fraudulent activities.

This skill provides a comprehensive framework for safeguarding referral programs against exploitation tactics like duplicate accounts, IP spoofing, and payout anomalies. It helps teams implement robust signal collection, risk scoring models, and automated investigation workflows to ensure program integrity while maintaining a positive experience for legitimate users. By integrating investigation runbooks and policy checklists, it bridges the gap between technical monitoring and operational compliance for marketing, finance, and legal teams.

主な機能

01Integration loops to refine heuristics based on manual audit findings

02Standardized investigation runbooks for evidence gathering and resolution

0329 GitHub stars

04Advanced risk scoring models for cohorts and high-volume referrers

05Workflow automation for flagging suspicious activity and reward pausing

06Automated signal collection for IP/device matching and velocity checks

ユースケース

01Reporting on program integrity and compliance for legal and finance departments

02Investigating suspicious spikes in referral payouts or account creations

03Designing security safeguards for new marketing referral initiatives

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add gtmagents/gtm-agents fraud-detection

For use in Claude.ai and ChatGPT

Download Skill